OpenAI recently launched a new social media app called Sora, which works like a TikTok for AI-generated videos. In less than 24 hours of its invite-only launch, the app is already filled with unsettling content, especially deepfakes of OpenAI CEO Sam Altman. Videos on Sora show Altman in strange and often disturbing scenarios. In one, he appears in a factory of pink pigs, asking, “Are my piggies enjoying their slop?” In another, he’s in a field of Pokemon, where he turns to the camera and says, “I hope Nintendo doesn’t sue us.”
Survey
✅ Thank you for completing the survey!
Other videos depict Altman serving Pikachu and Cartman at Starbucks, yelling at McDonald’s customers, and stealing GPUs from a store, reports TechCrunch. Even on X, users are sharing AI-generated videos of Altman made using the new app.
Sora has a feature called cameo, which lets users create AI versions of themselves by uploading biometric data. Users can then generate videos featuring their cameo, and control who can use it. Altman has made his cameo public, which explains why the feed is overflowing with videos of him interacting with Pokemon, SpongeBob, and other characters.
This seems like a deliberate move by Altman, perhaps intended to show that he doesn’t view his product as dangerous. Yet users are already using his cameo to challenge the ethics of the app itself.
While Sora is impressive in how realistically it portrays physics and environments, this realism also makes the deepfakes more believable, raising risks for misuse. Even though OpenAI emphasises safety, the app’s early content shows how easily AI-generated deepfakes can spiral out of control.